The artifact survives.
The doc. The PR. The decision.
NOCTEM LABS
Arachne is a governed inquiry workspace — a place where difficult thinking becomes durable, inspectable structure instead of disappearing into a transcript.
For the people carrying the world’s hardest problems, Arachne helps ideas evolve without losing the wisdom that gave them life.
Arachne is in active development. Its first public demonstration is being prepared.
THE PROBLEM
Ask a model a hard question and an answer arrives in seconds. Return next week and the reasoning is gone — the evidence, the objections, the paths you rejected, the reason you chose what you chose.
The doc. The PR. The decision.
Evidence. Objections. Rejected options.
of developers use or plan to use AI tools in development.
trust the accuracy of what those tools produce.
2025 Stack Overflow Developer Survey (n≈33,000). More developers actively distrust AI accuracy (46%) than trust it.
The missing layer is continuity — not another model.
THE PRODUCT
On the left, you think out loud. On the right, your question becomes a living organism — claims, evidence, contradictions, voids, dissent, and review state, held as structure you can point at.
Concept design for Arachne: a dialogue workbench beside a living organism of evidence, unresolved questions, dissent, and human selection and review.
You think out loud. Models contribute; none of them owns the answer.
Claims, evidence, voids, dissent, and review state become durable structure.
You choose the part that matters.
That choice changes what the next turn is allowed to use.
A refusal to accept amnesia as the price of progress.
HOW IT WORKS
FORM — a hard question becomes provisional structure.
SELECT — you choose the organism, branch, or void that matters.
EVOLVE — that choice changes what the next turn may use.
The dialogue changes the organism. The organism changes the dialogue.
As inquiry matures it can deepen, branch, or surface what remains unresolved. Work that survives review may later be crystallized, hardened, or seeded into a bounded capability. Those later stages are design direction, not current behaviour.
WHY IT COMPOUNDS
Providers improve, get replaced, or disappear. In Arachne a model's identity is provenance, never authority. Plural models can contribute and disagree, and the disagreement is kept. What a person actually reviewed persists underneath all of it — and improves the starting point for the next question.
The intelligence is replaceable. The accumulated understanding is not.
The foundation underneath it — evidence that stays attached to its sources, structure a person can inspect and contest, understanding that survives the tools that produced it — is built to carry further work in research, infrastructure, and capabilities we have not yet earned the right to describe. We will describe them when they exist.
WHO IT'S FOR
Arachne is for work that cannot be finished in one sitting, contained by one discipline, or carried by one mind.
— questions that evolve through competing explanations, experiment, and revision.
— systems whose decisions, constraints, and tradeoffs must survive implementation and handoff.
— where the reasoning matters as much as the result, and where the data cannot leave.
HOW WE BUILD
Important ideas should not become detached from the questions that gave them life.
What remains unknown should stay present rather than hidden behind confident language.
Rejected directions retain lessons that become valuable in a different moment.
Technology should expand what a person can see and reason about without quietly taking ownership of what matters.
We hold ourselves to the same standard. Nothing is recorded as verified until it has been observed directly, and findings that contradict our own plans are published rather than quietly dropped.
We are not trying to preserve everything. We are trying to preserve what makes continued understanding possible.

FOUNDER
2008 — Writing control logic for networked systems as a teenager, already wanting software that kept how a system got that way, not just what it currently was.
20 years — Enterprise APIs and distributed systems across healthcare, insurance, retail and logistics — Winsupply, Great American Insurance, Ulta Beauty, Exiger. Paid to keep large systems coherent long after everyone forgot why they were built.
2026 — Completed Deep Atlas applied ML training in May. Arachne's governed architecture followed by August.
Learn an unfamiliar system deeply enough to build with it. Then move the mechanism somewhere it has never been.
CURRENT STATUS
Noctem Labs is an early-stage, self-funded software research and development company. Arachne is in active development and its first public demonstration is being prepared.
The product is still being built. The mission is already clear.
Product language on this site describes the intended direction and values of Arachne. It is not a claim that every envisioned capability is complete or publicly available.